bstock_RFM_Analysis_demonstration
Umair
12/17/2018
Executive Summary
- Event data for ‘Bid Placed’ and ‘Auction Won’ events for the past 90 days was used in the analysis
- Users were segmented into five clusters for each of recency (last time event was registered from the user), and frequency (number of events registered from the user in the mentioned time-frame) by quintiles (20th percentiles).
Business Objective
Our goal is to optimize our marketing efforts by identifying high-value clients and segmenting our users by their recency and frequency of transactions to employ appropriate marketing campaigns for each segment. We will initially explore the events ‘Bid Placed’ and ‘Auction Won’ for the last 90 days and segment our users by their R and F. Following will be some examples about the kind of summaries we can get from this analysis.
Summary
- There are a total of 490 unique users in the data
- The following users appear in the top 20 users by each of their Recency and Frequency scores by both ‘Bid Placed’ and ‘Auction Won’ events:
## [1] "ajohnson.drph@gmail.com" "armendeshrefi@yahoo.com"
## [3] "info@airwavetelecom.ca" "james@victoryelitetech.com"
## [5] "lesfanbs@yahoo.com" "badmancustoms001@gmail.com"- We have 17 top frequency and recency users for the ‘Auction Won’ event.
- We have 43 top frequency and recency users for the ‘Bid Placed’ event.
Details: Bid Placed
Users in each recency and frequency bin
Distribution of frequency (how many times users triggered the event in the past 90 days) and recency (days since last event)
Contingency table for each frequency and recency bins
## frequencyBin
## recencyBin [ 1, 3) 3 [ 4, 7) [ 7, 14) [14,169]
## [54,60] 39 14 13 12 4
## [48,54) 34 13 22 19 14
## [42,48) 30 7 20 17 12
## [40,42) 22 5 17 16 19
## [35,40) 31 10 28 29 43As heatmap (red = low, white = high)
Each one of these segments differ from the others in their loyalty and frequency of transactions, so our marketing efforts can be customized for each. E.g users with the highest Recency and Frequency scores can be early adopters of new products, and could be rewarded, or upsold higher value products. Users with lesser scores can be potential loyalists, so they can be offered memberships, or loyalty programs, or recommended other products. Users with high Recency scores but low Frequency can be provided on-boarding support, or given early success. Others that had high frequency but low recency scores, meaning that it has been a while since they visited, but are high-value users, can be offered new products or exclusive offers to win them back.
High frequency users that haven’t placed bids since long
## [1] "acadianaltewireless@gmail.com" "cerhdz1987@gmail.com"
## [3] "misbawu45@hotmail.com" "mobilestaxx1@gmail.com"20 of the most high-value users with highest frequencies, as well as recent transactions
## email R_score F_score
## 1 ajohnson.drph@gmail.com 5 5
## 2 armendeshrefi@yahoo.com 5 5
## 3 arrg.inc@gmail.com 5 5
## 4 badmancustoms001@gmail.com 5 5
## 5 bmhonda2@gmail.com 5 5
## 6 caskellc@yahoo.com 5 5
## 7 charlesmorett@comcast.net 5 5
## 8 dmventuresllc1@gmail.com 5 5
## 9 doorsrusgambia@yahoo.com 5 5
## 10 getmemo@yahoo.com 5 5
## 11 gmpimport@yahoo.com 5 5
## 12 hallmark919@aol.com 5 5
## 13 hope_anaebo@yahoo.com 5 5
## 14 info@airwavetelecom.ca 5 5
## 15 james@victoryelitetech.com 5 5
## 16 k_ahoubim@yahoo.com 5 5
## 17 kamaljoe522@gmail.com 5 5
## 18 kennypost@icloud.com 5 5
## 19 ladmediagroup@gmail.com 5 5
## 20 lesfanbs@yahoo.com 5 5Trend of bid placements in our time-frame
hover over line for details
Auction Won
Users in each recency and frequency bin
Distribution of frequency (how many times users triggered the event in the past 90 days) and recency (days since last event)
Contingency table for users falling in each frequency and recency bins
## frequencyBin
## recencyBin 1 2 3 [4,26]
## [54,60] 39 13 1 2
## [51,54) 30 10 8 4
## [47,51) 18 7 9 3
## [41,47) 32 13 9 10
## [35,41) 29 16 9 17As heatmap
Users that won recently but only once in the 90-day span
## [1] "david@newtechwholesalers.com" "mobilestaxx1@gmail.com"20 of the most high-value users with highest frequencies, as well as recent transactions
## email R_score F_score
## 1 adamasfour86@gmail.com 5 4
## 2 ajohnson.drph@gmail.com 5 4
## 3 armendeshrefi@yahoo.com 5 4
## 4 indesigntrade@gmail.com 5 4
## 5 info@airwavetelecom.ca 5 4
## 6 james@victoryelitetech.com 5 4
## 7 lesfanbs@yahoo.com 5 4
## 8 m3mechanical@outlook.com 5 4
## 9 marcelo@solutekusa.com 5 4
## 10 mdposcom@gmail.com 5 4
## 11 pichottoraymond@gmail.com 5 4
## 12 rashidgauba@gmail.com 5 4
## 13 thephonedoctorva@gmail.com 5 4
## 14 trippj2017@outlook.com 5 4
## 15 tzippy15@gmail.com 5 4
## 16 vz@vizionz.net 5 4
## 17 wikiwoo@wikiwooelectronics.com 5 4
## 18 alidarwiche21@hotmail.com 5 3
## 19 badmancustoms001@gmail.com 5 3
## 20 changster7@gmail.com 5 3Trend of auction wins in our time-frame
